Recovery from DSM-IV post-traumatic stress disorder in the WHO World Mental Health surveys

BACKGROUND

Research on post-traumatic stress disorder (PTSD) course finds a substantial proportion of cases remit within 6 months, a majority within 2 years, and a substantial minority persists for many years. Results are inconsistent about pre-trauma predictors.

METHODS

The WHO World Mental Health surveys assessed lifetime DSM-IV PTSD presence-course after one randomly-selected trauma, allowing retrospective estimates of PTSD duration. Prior traumas, childhood adversities (CAs), and other lifetime DSM-IV mental disorders were examined as predictors using discrete-time person-month survival analysis among the 1575 respondents with lifetime PTSD.

RESULTS

20%, 27%, and 50% of cases recovered within 3, 6, and 24 months and 77% within 10 years (the longest duration allowing stable estimates). Time-related recall bias was found largely for recoveries after 24 months. Recovery was weakly related to most trauma types other than very low [odds-ratio (OR) 0.2-0.3] early-recovery (within 24 months) associated with purposefully injuring/torturing/killing and witnessing atrocities and very low later-recovery (25+ months) associated with being kidnapped. The significant ORs for prior traumas, CAs, and mental disorders were generally inconsistent between early- and later-recovery models. Cross-validated versions of final models nonetheless discriminated significantly between the 50% of respondents with highest and lowest predicted probabilities of both early-recovery (66-55% v. 43%) and later-recovery (75-68% v. 39%).

CONCLUSIONS

We found PTSD recovery trajectories similar to those in previous studies. The weak associations of pre-trauma factors with recovery, also consistent with previous studies, presumably are due to stronger influences of post-trauma factors.

First onset of suicidal thoughts and behaviours in college

BACKGROUND

College students are a worldwide increasing group of young people at risk for suicidal thoughts and behaviours (STB). However, no previous studies have prospectively investigated the first onset of STB during the college period.

METHODS

Using longitudinal data from the Leuven College Surveys, 2337 (response rate [RR]=66.6%) incoming freshmen provided baseline data on STB, parental psychopathology, childhood-adolescent traumatic experiences, 12-month risk for mental disorders, and 12-month stressful experiences. A total of 1253 baseline respondents provided data on 12-month STB in a two-year annual follow-up survey (conditional RR=53.6%; college dropout adjusted conditional RR=70.2%).

RESULTS

One-year incidence of first-onset STB was 4.8-6.4%. Effect sizes of the included risk factors varied considerably whether viewed from individual-level (ORs=1.91-17.58) or population-level perspective (PARPs=3.4-34.3%). Dating violence prior to the age of 17, physical abuse prior to the age of 17, and 12-month betrayal by someone else than the partner were most strong predictors for first-onset suicidal ideation (ORs=4.23-12.25; PARPs=8.7-27.1%) and plans (ORs=6.57-17.58; PARPs=15.2-34.3%). Multivariate prediction (AUC=0.84-0.91) revealed that 50.7-65.7% of first-onset STB cases were concentrated in the 10% at highest predicted risk.

LIMITATIONS

As this is a first investigation of STB onset in college, future studies should use validation samples to test the accuracy of our multivariate prediction model.

CONCLUSIONS

The first onset of STB in college appears to be higher than in the general population. Screening at college entrance is a promising strategy to identify those students at highest prospective risk, enabling the cost-efficient clinical assessment of young adults in college.

Mental disorders among college students in the World Health Organization World Mental Health Surveys

BACKGROUND

Although mental disorders are significant predictors of educational attainment throughout the entire educational career, most research on mental disorders among students has focused on the primary and secondary school years.

METHOD

The World Health Organization World Mental Health Surveys were used to examine the associations of mental disorders with college entry and attrition by comparing college students (n = 1572) and non-students in the same age range (18-22 years; n = 4178), including non-students who recently left college without graduating (n = 702) based on surveys in 21 countries (four low/lower-middle income, five upper-middle-income, one lower-middle or upper-middle at the times of two different surveys, and 11 high income). Lifetime and 12-month prevalence and age-of-onset of DSM-IV anxiety, mood, behavioral and substance disorders were assessed with the Composite International Diagnostic Interview (CIDI).

RESULTS

One-fifth (20.3%) of college students had 12-month DSM-IV/CIDI disorders; 83.1% of these cases had pre-matriculation onsets. Disorders with pre-matriculation onsets were more important than those with post-matriculation onsets in predicting subsequent college attrition, with substance disorders and, among women, major depression the most important such disorders. Only 16.4% of students with 12-month disorders received any 12-month healthcare treatment for their mental disorders.

CONCLUSIONS

Mental disorders are common among college students, have onsets that mostly occur prior to college entry, in the case of pre-matriculation disorders are associated with college attrition, and are typically untreated. Detection and effective treatment of these disorders early in the college career might reduce attrition and improve educational and psychosocial functioning.

Occupational differences in US Army suicide rates

BACKGROUND

Civilian suicide rates vary by occupation in ways related to occupational stress exposure. Comparable military research finds suicide rates elevated in combat arms occupations. However, no research has evaluated variation in this pattern by deployment history, the indicator of occupation stress widely considered responsible for the recent rise in the military suicide rate.

METHOD

The joint associations of Army occupation and deployment history in predicting suicides were analysed in an administrative dataset for the 729 337 male enlisted Regular Army soldiers in the US Army between 2004 and 2009.

RESULTS

There were 496 suicides over the study period (22.4/100 000 person-years). Only two occupational categories, both in combat arms, had significantly elevated suicide rates: infantrymen (37.2/100 000 person-years) and combat engineers (38.2/100 000 person-years). However, the suicide rates in these two categories were significantly lower when currently deployed (30.6/100 000 person-years) than never deployed or previously deployed (41.2-39.1/100 000 person-years), whereas the suicide rate of other soldiers was significantly higher when currently deployed and previously deployed (20.2-22.4/100 000 person-years) than never deployed (14.5/100 000 person-years), resulting in the adjusted suicide rate of infantrymen and combat engineers being most elevated when never deployed [odds ratio (OR) 2.9, 95% confidence interval (CI) 2.1-4.1], less so when previously deployed (OR 1.6, 95% CI 1.1-2.1), and not at all when currently deployed (OR 1.2, 95% CI 0.8-1.8). Adjustment for a differential 'healthy warrior effect' cannot explain this variation in the relative suicide rates of never-deployed infantrymen and combat engineers by deployment status.

CONCLUSIONS

Efforts are needed to elucidate the causal mechanisms underlying this interaction to guide preventive interventions for soldiers at high suicide risk.

Understanding the elevated suicide risk of female soldiers during deployments

BACKGROUND

The Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS) has found that the proportional elevation in the US Army enlisted soldier suicide rate during deployment (compared with the never-deployed or previously deployed) is significantly higher among women than men, raising the possibility of gender differences in the adverse psychological effects of deployment.

METHOD

Person-month survival models based on a consolidated administrative database for active duty enlisted Regular Army soldiers in 2004-2009 (n = 975,057) were used to characterize the gender × deployment interaction predicting suicide. Four explanatory hypotheses were explored involving the proportion of females in each soldier's occupation, the proportion of same-gender soldiers in each soldier's unit, whether the soldier reported sexual assault victimization in the previous 12 months, and the soldier's pre-deployment history of treated mental/behavioral disorders.

RESULTS

The suicide rate of currently deployed women (14.0/100,000 person-years) was 3.1-3.5 times the rates of other (i.e. never-deployed/previously deployed) women. The suicide rate of currently deployed men (22.6/100,000 person-years) was 0.9-1.2 times the rates of other men. The adjusted (for time trends, sociodemographics, and Army career variables) female:male odds ratio comparing the suicide rates of currently deployed v. other women v. men was 2.8 (95% confidence interval 1.1-6.8), became 2.4 after excluding soldiers with Direct Combat Arms occupations, and remained elevated (in the range 1.9-2.8) after adjusting for the hypothesized explanatory variables.

CONCLUSIONS

These results are valuable in excluding otherwise plausible hypotheses for the elevated suicide rate of deployed women and point to the importance of expanding future research on the psychological challenges of deployment for women.

The effects of temporally secondary co-morbid mental disorders on the associations of DSM-IV ADHD with adverse outcomes in the US National Comorbidity Survey Replication Adolescent Supplement (NCS-A)

BACKGROUND

Although DSM-IV attention deficit hyperactivity disorder (ADHD) is known to be associated with numerous adverse outcomes, uncertainties exist about how much these associations are mediated temporally by secondary co-morbid disorders.

METHOD

The US National Comorbidity Survey Replication Adolescent Supplement (NCS-A), a national survey of adolescents aged 13-17 years (n = 6483 adolescent-parent pairs), assessed DSM-IV disorders with the World Health Organization (WHO) Composite International Diagnostic Interview (CIDI). Statistical decomposition was used to compare direct effects of ADHD with indirect effects of ADHD through temporally secondary mental disorders (anxiety, mood, disruptive behavior, substance disorders) in predicting poor educational performance (suspension, repeating a grade, below-average grades), suicidality (ideation, plans, attempts) and parent perceptions of adolescent functioning (physical and mental health, interference with role functioning and distress due to emotional problems).

RESULTS

ADHD had significant gross associations with all outcomes. Direct effects of ADHD explained most (51.9-67.6%) of these associations with repeating a grade in school, perceived physical and mental health (only girls), interference with role functioning and distress, and significant components (34.5-44.6%) of the associations with school suspension and perceived mental health (only boys). Indirect effects of ADHD on educational outcomes were predominantly through disruptive behavior disorders (26.9-52.5%) whereas indirect effects on suicidality were predominantly through mood disorders (42.8-59.1%). Indirect effects on most other outcomes were through both mood (19.8-31.2%) and disruptive behavior (20.1-24.5%) disorders, with anxiety and substance disorders less consistently important. Most associations were comparable for girls and boys.

CONCLUSIONS

Interventions aimed at reducing the adverse effects of ADHD might profitably target prevention or treatment of temporally secondary co-morbid disorders.

Lifetime co-morbidity of DSM-IV disorders in the US National Comorbidity Survey Replication Adolescent Supplement (NCS-A)

BACKGROUND

Research on the structure of co-morbidity among common mental disorders has largely focused on current prevalence rather than on the development of co-morbidity. This report presents preliminary results of the latter type of analysis based on the US National Comorbidity Survey Replication Adolescent Supplement (NCS-A).

METHOD

A national survey was carried out of adolescent mental disorders. DSM-IV diagnoses were based on the Composite International Diagnostic Interview (CIDI) administered to adolescents and questionnaires self-administered to parents. Factor analysis examined co-morbidity among 15 lifetime DSM-IV disorders. Discrete-time survival analysis was used to predict first onset of each disorder from information about prior history of the other 14 disorders.

RESULTS

Factor analysis found four factors representing fear, distress, behavior and substance disorders. Associations of temporally primary disorders with the subsequent onset of other disorders, dated using retrospective age-of-onset (AOO) reports, were almost entirely positive. Within-class associations (e.g. distress disorders predicting subsequent onset of other distress disorders) were more consistently significant (63.2%) than between-class associations (33.0%). Strength of associations decreased as co-morbidity among disorders increased. The percentage of lifetime disorders explained (in a predictive rather than a causal sense) by temporally prior disorders was in the range 3.7-6.9% for earliest-onset disorders [specific phobia and attention deficit hyperactivity disorder (ADHD)] and much higher (23.1-64.3%) for later-onset disorders. Fear disorders were the strongest predictors of most other subsequent disorders.

CONCLUSIONS

Adolescent mental disorders are highly co-morbid. The strong associations of temporally primary fear disorders with many other later-onset disorders suggest that fear disorders might be promising targets for early interventions.

Childhood adversities and adult psychopathology in the National Comorbidity Survey Replication (NCS-R) III: associations with functional impairment related to DSM-IV disorders

BACKGROUND

Despite evidence that childhood adversities (CAs) are associated with increased risk of mental disorders, little is known about their associations with disorder-related impairment. We report the associations between CAs and functional impairment associated with 12-month DSM-IV disorders in a national sample.

METHOD

We used data from the US National Comorbidity Survey Replication (NCS-R). Respondents completed diagnostic interviews that assessed 12-month DSM-IV disorder prevalence and impairment. Associations of 12 retrospectively reported CAs with impairment among cases (n=2242) were assessed using multiple regression analysis. Impairment measures included a dichotomous measure of classification in the severe range of impairment on the Sheehan Disability Scale (SDS) and a measure of self-reported number of days out of role due to emotional problems in the past 12 months.

RESULTS

CAs were positively and significantly associated with impairment. Predictive effects of CAs on the SDS were particularly pronounced for anxiety disorders and were significant in predicting increased days out of role associated with mood, anxiety and disruptive behavior disorders. Predictive effects persisted throughout the life course and were not accounted for by disorder co-morbidity. CAs associated with maladaptive family functioning (MFF; parental mental illness, substance disorder, criminality, family violence, abuse, neglect) were more consistently associated with impairment than other CAs. The joint effects of co-morbid MFF CAs were significantly subadditive. Simulations suggest that CAs account for 19.6% of severely impairing disorders and 17.4% of days out of role.

CONCLUSIONS

CAs predict greater disorder-related impairment, highlighting the ongoing clinical significance of CAs at every stage of the life course.

Multiple imputation in a large-scale complex survey: a practical guide

The Cancer Care Outcomes Research and Surveillance (CanCORS) Consortium is a multisite, multimode, multiwave study of the quality and patterns of care delivered to population-based cohorts of newly diagnosed patients with lung and colorectal cancer. As is typical in observational studies, missing data are a serious concern for CanCORS, following complicated patterns that impose severe challenges to the consortium investigators. Despite the popularity of multiple imputation of missing data, its acceptance and application still lag in large-scale studies with complicated data sets such as CanCORS. We use sequential regression multiple imputation, implemented in public-available software, to deal with non-response in the CanCORS surveys and construct a centralised completed database that can be easily used by investigators from multiple sites. Our work illustrates the feasibility of multiple imputation in a large-scale multiobjective survey, showing its capacity to handle complex missing data. We present the implementation process in detail as an example for practitioners and discuss some of the challenging issues which need further research.

Education and race-ethnicity differences in the lifetime risk of alcohol dependence

OBJECTIVES

While lower socioeconomic status (SES) is related to higher risk for alcohol dependence, minority race-ethnicity is often associated with lower risk. This study attempts to clarify the nature and extent of social inequalities in alcohol dependence by investigating the effects of SES and race-ethnicity on the development of alcohol dependence following first alcohol use.

DESIGN

Cross-sectional analysis of data from the National Epidemiologic Survey on Alcohol and Related Conditions (n = 43,093). Survival analysis was used to model alcohol dependence onset according to education, race-ethnicity and their interaction.

SETTING

United States, 2001-2.

RESULTS

Compared with non-Hispanic white people, age-adjusted and sex-adjusted risks of alcohol dependence were lower among black people (odds ratio (OR) = 0.70, 95% confidence interval (CI) = 0.63 to 0.78), Asians (OR = 0.65, CI = 0.49 to 0.86) and Hispanics (OR = 0.68, CI = 0.58 to 0.79) and higher among American Indians (OR = 1.37, CI = 1.09 to 1.73). Individuals without a college degree had higher risks of alcohol dependence than individuals with a college degree or more; however, the magnitude of risk varied significantly by race-ethnicity (chi(2) for the interaction between education and race-ethnicity = 19.7, df = 10, p = 0.03); odds ratios for less than a college degree were 1.12, 1.46, 2.24, 2.35 and 10.99 among Hispanics, white people, black people, Asians, and American Indians, respectively. There was no association between education and alcohol dependence among Hispanics.

CONCLUSIONS

Race-ethnicity differences in the magnitude of the association between education and alcohol dependence suggest that aspects of racial-ethnic group membership mitigate or exacerbate the effects of social adversity.

Cross-national prevalence and correlates of adult attention-deficit hyperactivity disorder

BACKGROUND

Little is known about the epidemiology of adult attention-deficit hyperactivity disorder (ADHD).

AIMS

To estimate the prevalence and correlates of DSM-IV adult ADHD in the World Health Organization World Mental Health Survey Initiative.

METHOD

An ADHD screen was administered to respondents aged 18-44 years in ten countries in the Americas, Europe and the Middle East (n=11422). Masked clinical reappraisal interviews were administered to 154 US respondents to calibrate the screen. Multiple imputation was used to estimate prevalence and correlates based on the assumption of cross-national calibration comparability.

RESULTS

Estimates of ADHD prevalence averaged 3.4% (range 1.2-7.3%), with lower prevalence in lower-income countries (1.9%) compared with higher-income countries (4.2%). Adult ADHD often co-occurs with other DSM-IV disorders and is associated with considerable role disability. Few cases are treated for ADHD, but in many cases treatment is given for comorbid disorders.

CONCLUSIONS

Adult ADHD should be considered more seriously in future epidemiological and clinical studies than is currently the case.

Recalibration methods to enhance information on prevalence rates from large mental health surveys

Comparisons between self-report and clinical psychiatric measures have revealed considerable disagreement. It is unsafe to consider these measures as directly equivalent, so it would be valuable to have a reliable recalibration of one measure in terms of the other. We evaluated multiple imputation incorporating a Bayesian approach, and a fully Bayesian method, to recalibrate diagnoses from a self-report survey interview in terms of those from a clinical interview with data from a two-phase national household survey for a practical application, and artificial data for simulation studies. The most important factors in obtaining a precise and accurate 'clinical' prevalence estimate from self-report data were (a) good agreement between the two diagnostic measures and (b) a sufficiently large set of calibration data with diagnoses based on both kinds of interview from the same group of subjects. From the case study, calibration data on 612 subjects were sufficient to yield estimates of the total prevalence of anxiety, depression or neurosis with a precision in the region of +/-2%. The limitations of the calibration method demonstrate the need to increase agreement between survey and reference measures by improving lay interviews and their diagnostic algorithms.

Short screening scales to monitor population prevalences and trends in non-specific psychological distress

BACKGROUND

A 10-question screening scale of psychological distress and a six-question short-form scale embedded within the 10-question scale were developed for the redesigned US National Health Interview Survey (NHIS).

METHODS

Initial pilot questions were administered in a US national mail survey (N = 1401). A reduced set of questions was subsequently administered in a US national telephone survey (N = 1574). The 10-question and six-question scales, which we refer to as the K10 and K6, were constructed from the reduced set of questions based on Item Response Theory models. The scales were subsequently validated in a two-stage clinical reappraisal survey (N = 1000 telephone screening interviews in the first stage followed by N = 153 face-to-face clinical interviews in the second stage that oversampled first-stage respondents who screened positive for emotional problems) in a local convenience sample. The second-stage sample was administered the screening scales along with the Structured Clinical Interview for DSM-IV (SCID). The K6 was subsequently included in the 1997 (N = 36116) and 1998 (N = 32440) US National Health Interview Survey, while the K10 was included in the 1997 (N = 10641) Australian National Survey of Mental Health and Well-Being.

RESULTS

Both the K10 and K6 have good precision in the 90th-99th percentile range of the population distribution (standard errors of standardized scores in the range 0.20-0.25) as well as consistent psychometric properties across major sociodemographic subsamples. The scales strongly discriminate between community cases and non-cases of DSM-IV/SCID disorders, with areas under the Receiver Operating Characteristic (ROC) curve of 0.87-0.88 for disorders having Global Assessment of Functioning (GAF) scores of 0-70 and 0.95-0.96 for disorders having GAF scores of 0-50.

CONCLUSIONS

The brevity, strong psychometric properties, and ability to discriminate DSM-IV cases from non-cases make the K10 and K6 attractive for use in general-purpose health surveys. The scales are already being used in annual government health surveys in the US and Canada as well as in the WHO World Mental Health Surveys. Routine inclusion of either the K10 or K6 in clinical studies would create an important, and heretofore missing, crosswalk between community and clinical epidemiology.

Statistical issues in reporting quality data: small samples and casemix variation

PURPOSE

To present two key statistical issues that arise in analysis and reporting of quality data.

SUMMARY

Casemix variation is relevant to quality reporting when the units being measured have differing distributions of patient characteristics that also affect the quality outcome. When this is the case, adjustment using stratification or regression may be appropriate. Such adjustments may be controversial when the patient characteristic does not have an obvious relationship to the outcome. Stratified reporting poses problems for sample size and reporting format, but may be useful when casemix effects vary across units. Although there are no absolute standards of reliability, high reliabilities (interunit F > or = 10 or reliability > or = 0.9) are desirable for distinguishing above- and below-average units. When small or unequal sample sizes complicate reporting, precision may be improved using indirect estimation techniques that incorporate auxiliary information, and 'shrinkage' estimation can help to summarize the strength of evidence about units with small samples.

CONCLUSIONS

With broader understanding of casemix adjustment and methods for analyzing small samples, quality data can be analysed and reported more accurately.

National quality monitoring of Medicare health plans: the relationship between enrollees' reports and the quality of clinical care

BACKGROUND

The clinical quality of health plans varies. The associations between different measures of health plan quality are incompletely understood.

OBJECTIVE

To assess the relationships between enrollee reports on the quality of health plans as measured by the Consumer Assessment of Health Plans Study (CAHPS 2.0) survey and the clinical quality of care measured by the Medicare Health Plan Employer Data and Information Set (HEDIS).

DESIGN

Observational cohort study.

SAMPLE

National sample of 233 Medicare health plans that reported data using the CAHPS 2.0 survey and Medicare HEDIS during 1998.

MEASURES

Five composite measures and four ratings derived from the CAHPS survey and six measures of clinical quality from Medicare HEDIS.

RESULTS

Two composite measures ("getting needed care" and "health plan information and customer service") were significantly associated with most of the HEDIS clinical quality measures. The proportion of enrollees having a personal doctor was also significantly associated with rates of mammography, eye exams for diabetics, beta-blocker use after myocardial infarction, and follow-up after mental health hospitalization. Enrollees' ratings of health plan care were less consistently associated with HEDIS performance. In multivariable analyses, the measure of health plan communication ("health plan information and customer service") was the most consistent predictor of HEDIS performance.

CONCLUSIONS

The pattern of associations we observed among some of the measures suggests that the CAHPS survey and HEDIS are complementary quality monitoring strategies. Our results suggest that health plans that provide better access and customer service also provide better clinical care.

Racial disparity in influenza vaccination: does managed care narrow the gap between African Americans and whites?

CONTEXT

Substantial racial disparities exist in use of some health services. Whether managed care could reduce racial disparities in the use of preventive services is not known.

OBJECTIVE

To determine whether the magnitude of racial disparity in influenza vaccination is smaller among managed care enrollees than among those with fee-for-service insurance.

DESIGN, SETTING, AND PARTICIPANTS

The 1996 Medicare Current Beneficiary Survey of a US cohort of 13 674 African American and white Medicare beneficiaries with managed care and fee-for-service insurance.

MAIN OUTCOME MEASURES

Percentage of respondents (adjusted for sociodemographic characteristics, clinical comorbid conditions, and care-seeking attitudes) who received influenza vaccination and magnitude of racial disparity in influenza vaccination, compared among those with managed care and fee-for-service insurance.

RESULTS

Eight percent of the beneficiaries were African American and 11% were enrolled in managed care. Overall, 65.8% received influenza vaccination. Whites were substantially more likely to be vaccinated than African Americans (67.7% vs 46.1%; absolute disparity, 21.6%; 95% confidence interval [CI], 18.2%-25.0%). Managed care enrollees were more likely than those with fee-for-service insurance to receive influenza vaccination (71.2% vs 65.4%; difference, 5.8%; 95% CI, 3.6%-8.3%). The adjusted racial disparity in fee-for-service was 24.9% (95% CI, 19.6%-30.1%) and in managed care was 18.6% (95% CI, 9.8%-27.4%). These adjusted racial disparities were both statistically significant, but the absolute percentage point difference in racial disparity between the 2 insurance groups (6.3%; 95% CI, -4.6% to 17.2%) was not.

CONCLUSION

Managed care is associated with higher rates of influenza vaccination for both whites and African Americans, but racial disparity in vaccination is not reduced in managed care. Our results suggest that additional efforts are needed to adequately address this disparity.

Physician incentives and disclosure of payment methods to patients

OBJECTIVE

There is increasing public discussion of the value of disclosing how physicians are paid. However, little is known about patients' awareness of and interest in physician payment information or its potential impact on patients' evaluation of their care.

DESIGN

Cross-sectional survey

SETTING

Managed care and indemnity plans of a large, national health insurer.

PARTICIPANTS

Telephone interviews were conducted with 2,086 adult patients in Atlanta, Ga; Baltimore, Md/Washington DC; and Orlando, Fla (response rate, 54%).

MEASUREMENTS AND MAIN RESULTS

Patients were interviewed to assess perceptions of their physicians' payment method, preference for disclosure, and perceived effect of different financial incentives on quality of care. Non-managed fee-for-service patients (44%) were more likely to correctly identify how their physicians were paid than those with salaried (32%) or capitated (16%) physicians. Just over half (54%) wanted to be informed about their physicians' payment

METHOD

Patients of capitated and salaried physicians were as likely to want disclosure as patients of fee-for-service physicians. College graduates were more likely to prefer disclosure than other patients. Many patients (76%) thought a bonus paid for ordering fewer than the average number of tests would adversely affect the quality of their care. About half of the patients (53%) thought a particular type of withhold would adversely affect the quality of their care. White patients, college graduates, and those who had higher incomes were more likely to think that these types of bonuses and withholds would have a negative impact on their care. Among patients who believed that these types of bonuses adversely affected care, those with non-managed fee-for-service insurance and college graduates were more willing to pay a higher deductible or co-payment in order to get tests that they thought were necessary.

CONCLUSIONS

Most patients were unaware of how their physicians are paid, and only about half wanted to know. Most believed that bonuses or withholds designed to reduce the use of services would adversely affect the quality of their care. Lack of knowledge combined with strong attitudes about various financial incentives suggest that improved patient education could clarify patient understanding of the nature and rationale for different types of incentives. More public discussion of this important topic is warranted.

Health plan characteristics and consumers' assessments of quality

Many purchasers and consumers of health care have become concerned about the quality of care being delivered in managed care plans. Little is known, however, about the health plan characteristics that are associated with better performance. We used survey responses from 82,583 Medicare beneficiaries from 182 health plans to study the association of consumers' assessments of care with health plan characteristics. For-profit and nationally affiliated health plans received much worse scores on the outcomes of interest, particularly for overall ratings of the health plan and composite measures of customer service and access to care. Health plans accredited by the National Committee for Quality Assurance did not receive higher scores.

Unmet health needs of uninsured adults in the United States

CONTEXT

In 1998, 33 million US adults aged 18 to 64 years lacked health insurance. Determining the unmet health needs of this population may aid efforts to improve access to care.

OBJECTIVE

To compare nationally representative estimates of the unmet health needs of uninsured and insured adults, particularly among persons with major health risks.

DESIGN AND SETTING

Random household telephone survey conducted in all 50 states and the District of Columbia through the Behavioral Risk Factor Surveillance System.

PARTICIPANTS

A total of 105,764 adults aged 18 to 64 years in 1997 and 117,364 in 1998, classified as long-term (>/=1 year) uninsured (9.7%), short-term (<1 year) uninsured (4.3%), or insured (86.0%).

MAIN OUTCOME MEASURES

Adjusted proportions of participants who could not see a physician when needed due to cost in the past year, had not had a routine checkup within 2 years, and had not received clinically indicated preventive services, compared by insurance status.

RESULTS

Long-term- and short-term-uninsured adults were more likely than insured adults to report that they could not see a physician when needed due to cost (26.8%, 21.7%, and 8.2%, respectively), especially among those in poor health (69.1%, 51.9%, and 21.8%) or fair health (48.8%, 42.4%, and 15.7%) (P<.001). Long-term-uninsured adults in general were much more likely than short-term-uninsured and insured adults not to have had a routine checkup in the last 2 years (42.8%, 22.3%, and 17.8%, respectively) and among smokers, obese individuals, binge drinkers, and people with hypertension, elevated cholesterol, diabetes, or human immunodeficiency virus risk factors (P<.001). Deficits in cancer screening, cardiovascular risk reduction, and diabetes care were most pronounced among long-term-uninsured adults.

CONCLUSIONS

In our study, long-term-uninsured adults reported much greater unmet health needs than insured adults. Providing insurance to improve access to care for long-term-uninsured adults, particularly those with major health risks, could have substantial clinical benefits. JAMA. 2000;284:2061-2069

Impact of sociodemographic case mix on the HEDIS measures of health plan quality

BACKGROUND

The widely used Health Plan Employer Data and Information Set (HEDIS) measures may be affected by differences among plans in sociodemographic characteristics of members.

OBJECTIVE

The objective of this study was to estimate effects of geographically linked patient sociodemographic characteristics on differential performance within and among plans on HEDIS measures.

RESEARCH DESIGN

Using logistic regression, we modeled associations between age, sex, and residential area characteristics of health plan members and results on HEDIS measures. We then calculated the impact of adjusting for these associations on plan-level measures.

SUBJECTS

This study included 92,232 commercially insured members with individual-level HEDIS data and an additional 20,615 members whose geographic distribution was provided.

MEASURES

This study used 7 measures of screening and preventive services.

RESULTS

Performance was negatively associated with percent receiving public assistance in the local area (6 of 7 measures), percent black (5 measures), and percent Hispanic (2 measures) and positively associated with percent college educated (6 measures), percent urban (2 measures), and percent Asian (1 measure) after controlling for plan and product type. These effects were generally consistent across plans. When measures were adjusted for these characteristics, rates for most plans changed by less than 5 percentage points. The largest change in the difference between plans ranged from 1.5% for retinal exams for people with diabetes to 20.2% for immunization of adolescents.

CONCLUSIONS

Performance on quality indicators for individual members is associated with sociodemographic context. Adjustment has little impact on the measured performance of most plans but a substantial impact on a few. Further study with more plans is required to determine the appropriateness and feasibility of adjustment.

How consumer assessments of managed care vary within and among markets

This study investigated the extent to which the Consumer Assessments of Health Plans Survey (CAHPS) distinguishes performance of Medicare managed care (MMC) health plans. Results indicate that CAHPS ratings and report composites distinguish among plans both nationally and within markets. Global ratings of a health plan and reports on customer services varied strongly at the individual plan level, with smaller effects seen at regional and Metropolitan Statistical Area (MSA) levels. Ratings of doctors and health care, and reports on experiences in the doctor's office varied more among regions and among MSAs than among plans within the same MSA. These patterns are consonant with our hypotheses about the determinants of these ratings: customer service is a distinct plan function, but medical services are provided by networks that often overlap for plans in the same area. We conclude that the CAHPS-MMC survey can inform consumers choosing among plans as well as policymakers and researchers.

Does the effect of respondent characteristics on consumer assessments vary across health plans?

Responses to the Consumer Assessments of Health Plans Survey (CAHPS) are related to respondent characteristics. CAHPS procedures include casemix adjustment to remove effects of difference in respondent characteristics on comparative plan ratings, under the assumption that casemix coefficients are homogeneous across plans. The authors analyzed data from the Washington state CAHPS demonstration, fitting hierarchical models in which coefficients of casemix variables and intercepts could vary by plan. They estimated the impact of variability in casemix coefficients on plan adjustments and also assessed the implications for differential effects of individual characteristics at different plans. Estimated between-plan variability of coefficients was small, but the data are consistent with substantially larger variability. The potential impact of this variability on adjustments for plans was small relative to the magnitude of the adjustments. Comparisons between plans for individuals, however, could be affected substantially. This methodology could be useful wherever casemix adjustment is applied.

Physicians' reports of focused expertise in clinical practice

Little is known about the prevalence of focused expertise (special areas of expertise within a clinical field) among physicians, yet such expertise may influence how care is delivered. We surveyed general internists, pediatricians, cardiologists, infectious disease specialists, and orthopedic surgeons to describe the prevalence of focused expertise and identify associated physician and practice characteristics. About one quarter of generalists and three quarters of specialists reported a focused expertise within their primary specialty. Hospital-based physicians more often reported such expertise, and physicians reimbursed by capitation less often reported expertise. Learning how focused expertise affects processes and outcomes of care will contribute to decisions about physician training and staffing of medical groups.

Sex differences in assessments of the quality of Medicare managed care

Women rate their care slightly more positively than men on the Medicare managed care Consumer Assessments of Health Plans Study (CAHPS) survey. On four of five composites, women have comparable or slightly more positive composite scores than men. Responses to individual questions in 1997 indicate that women may have slightly more problems getting referrals, equipment, and assistance. In 1998, there were no differences in difficulty getting a referral or assistance, but women were less likely to say their plan provided help, services, and equipment. Women were less likely to get a flu shot in both years. Further monitoring of CAHPS data on sex differences in difficulty getting needed medical equipment are warranted.

Dimensions of consumer-assessed quality of Medicare managed-care health plans

OBJECTIVES

We investigated relationships at the health-plan level among member ratings of and reports on plans in the Consumer Assessment of Health Plans Survey (CAHPS). We sought a more parsimonious description of the reports that can be used in analyses of the distribution and correlates of consumer-assessed quality.

SUBJECTS

There were 89,419 Medicare beneficiaries enrolled in 212 Medicare managed-care health plans who responded to CAHPS in 1998.

MEASURES

There were 39 survey items measuring consumer ratings of and reports on care.

METHODS

We adjusted correlations for sampling variability in the plan means and performed a principal factor analysis of the report items with oblique rotation. We grouped items that loaded heavily on the different factors, formed composites, and regressed rating items on the report composites.

RESULTS

Four factors explained 75% of the variance in the reports. The corresponding groups of items were concerned with the following subjects: (1) interactions around delivery of care in the doctor's office; (2) customer service from the plan; (3) access to medical services provided by the plan, such as specialist care, equipment, therapy, or drugs; and (4) advice on health-promoting activities. Corrected Cronbach alpha for composites were 0.97, 0.93, 0.86, and 0.60. The "delivery" composite was strongly predictive of overall ratings of care, doctor, and specialist; the "customer" composite was strongly predictive of overall ratings of the plan.

CONCLUSIONS

CAHPS distinguishes among dimensions of between-plan variability of consumer-assessed quality. Different global ratings are related to distinct groups of consumer reports on their experiences.

Applying the Cox proportional hazards model for analysis of latency data with interval censoring

The latency time of an infectious disease is defined as the time from infection to disease onset. This paper applies the proportional hazards model to estimate the effect of covariates on latency when the time of disease onset is exact or right-censored but the time of infection is interval-censored. We use a Monte Carlo EM algorithm to estimate parameters of the joint distribution of infection times and latency times. At each EM iteration, exact infection times are multiply imputed from the density determined by the parameters of the infection and latency time distributions. The methodology is tested using a simulation study and is applied to data from a cohort of haemophiliacs with HIV disease.

Applying the Cox proportional hazards model when the change time of a binary time-varying covariate is interval censored

This paper develops methodology for estimation of the effect of a binary time-varying covariate on failure times when the change time of the covariate is interval censored. The motivating example is a study of cytomegalovirus (CMV) disease in patients with human immunodeficiency virus (HIV) disease. We are interested in determining whether CMV shedding predicts an increased hazard for developing active CMV disease. Since a clinical screening test is needed to detect CMV shedding, the time that shedding begins is only known to lie in an interval bounded by the patient's last negative and first positive tests. In a Cox proportional hazards model with a time-varying covariate for CMV shedding, the partial likelihood depends on the covariate status of every individual in the risk set at each failure time. Due to interval censoring, this is not always known. To solve this problem, we use a Monte Carlo EM algorithm with a Gibbs sampler embedded in the E-step. We generate multiple completed data sets by drawing imputed exact shedding times based on the joint likelihood of the shedding times and event times under the Cox model. The method is evaluated using a simulation study and is applied to the data set described above.

Physician experiences with, and ratings of, managed care organizations in Massachusetts

BACKGROUND

Physicians can provide important information about how managed care plans affect the delivery of health care. Assessments of the quality of managed care plans have rarely used physician evaluations.

OBJECTIVES

To elicit physician evaluations of managed care plans and to determine factors associated with those evaluations.

RESEARCH DESIGN

Physicians were asked in a mail survey to evaluate a managed care plan they were associated with.

SUBJECTS

Probability sample of 1,336 physicians associated with the five largest managed care health plans in Massachusetts.

MEASURES

Physicians were asked about the extent to which the management strategies used by a plan influenced their clinical behavior and about the quality of care available to their patients.

RESULTS

Evaluations of the plans were significantly different among the eight units evaluated. Some differences between divisions within plans were as large as differences among plans. Physicians reported that the use of education and peer influence influenced their clinical behavior and facilitated the provision of high quality care more than did rules and regulations or financial incentives. Physicians evaluated most positively plans, which they said used educational strategies more than other plans and which used rules and regulations and financial incentives less. Physicians tended to rate staff and group model plans more positively than did other plans.

CONCLUSIONS

Physicians can provide important information about the extent to which the organization and operation of managed care plans affect the provision of high quality care.

Sample allocation for overlapping domains in a physician survey with a limited population

We consider an analytic survey in which each survey unit can respond with respect to one or more domains to which it belongs. In this situation, the optimal sample allocation is constrained by the number of available units in each stratum defined by a set of domains. We present optimal sample allocation rules for this situation, and other rules that apply when there are additional constraints on sample sizes or variances by domain. We apply these rules to our motivating example, the design of a survey of physicians on their experiences with health plans, in which each physician can only be asked about her experience with one plan regardless of her number of affiliations.

Use of hormone replacement therapy by postmenopausal women in the United States

BACKGROUND

The benefits and risks of hormone replacement therapy (HRT) in postmenopausal women are not fully defined, and individual characteristics and preferences may influence decisions to use this therapy. Previous studies of postmenopausal women who use HRT have been conducted in local or highly selected cohorts or have not focused on current use.

OBJECTIVE

To examine sociodemographic, clinical, and psychological factors associated with current use of HRT in a national population-based cohort.

DESIGN

Random-digit telephone survey.

SETTING

Probability sample of U.S. households with a telephone.

PARTICIPANTS

495 postmenopausal women 50 to 74 years of age in 1995.

MEASUREMENTS

Current use of HRT.

RESULTS

Current use of HRT was reported by 37.6% of women (58.7% of those who underwent hysterectomy and 19.6% of those who did not undergo hysterectomy; P = 0.001). In multivariable analyses, use of HRT was more common among women in the South (adjusted odds ratio, 2.67 [95% CI, 1.08 to 6.59]) and West (odds ratio, 2.76 [CI, 1.01 to 7.53]) than the Northeast. Use was more common among college graduates (odds ratio, 3.72 [CI, 1.29 to 10.71]) and less common among women with diabetes mellitus (odds ratio, 0.17 [CI, 0.05 to 0.51]). Other cardiac risk factors and most psychological characteristics were not associated with HRT use.

CONCLUSIONS

Sociodemographic factors, such as region and education, may be more strongly associated with use of HRT than clinical factors, such as risk for cardiovascular disease. Future efforts should focus on understanding sociodemographic variations, defining which women are most likely to benefit, and targeting therapy to them.

The impact of having parents report about both their own and their children's experiences with health insurance plans

OBJECTIVES

The aim of this study was to determine whether parents rate their children's care differently when they also rate their own care than when they do not.

METHODS

Subjects were employees of Washington State who had been enrolled in a health plan for at least 6 months and who had at least one covered child. Subjects were randomly assigned to four study groups that were surveyed using different protocols. To assess the stability of responses over time, a follow-up telephone interview was conducted with individuals in two of the groups.

RESULTS

Parents or guardians who received both the Adult and Child Surveys were less likely to complete a survey than those who received only one survey. Responses to selected survey questions were quite stable between survey administrations. Parents who rated only their child's health care experiences generally gave more positive responses than those who also rated their own care, although few of these differences were statistically significant. This may have been due, in part, to the lower response rates in the latter group. The pairs of survey questions that ask about the adult's and child's experiences with the same aspects of care had moderate to high levels of association. The pair with the weakest association asked how clearly the doctor or nurse explained things to the adult or the child.

CONCLUSIONS

Sending both an adult and child survey to an adult could have an effect on the pattern of responses and result in lower response rates, but this might be a cost-effective way to collect reports about both adult and child health care.

A Markov chain Monte Carlo EM algorithm for analyzing interval-censored data under the Cox proportional hazards model

This paper proposes a Monte Carlo EM (MCEM) algorithm for fitting the proportional hazards model for interval-censored failure-time data. The algorithm generates orderings of the failures from their probability distribution under the model. We maximize the average of the log-likelihoods from these completed data sets to obtain updated parameter estimates. As with the standard Cox model, this algorithm does not require the estimation of the baseline hazard function. The performance of the algorithm is evaluated using simulations, and the method is applied to data from AIDS and cancer studies. Our results indicate that our method produced more precise and unbiased estimates than methods of right and midpoint imputation.

The relationship between method of physician payment and patient trust

CONTEXT

Trust is the cornerstone of the patient-physician relationship. Payment methods that place physicians at financial risk have raised concerns about patients' trust in physicians to act in patients' best interests.

OBJECTIVE

To evaluate the extent to which methods of physician payment are related to patient trust.

DESIGN

Cross-sectional telephone interview survey done between January and June 1997.

SETTING

Health plans of a large national insurer in Atlanta, Ga, the Baltimore, Md-Washington, DC, area, and Orlando, Fla.

PARTICIPANTS

A total of 2086 adult managed care and indemnity patients.

MAIN OUTCOME MEASURE

A 10-item scale (alpha = .94) assessing patients' trust in physicians.

RESULTS

More fee-for-service (FFS) indemnity patients (94%) completely or mostly trust their physicians to "put their health and well-being above keeping down the health plan's costs" than salary (77%), capitated (83%), or FFS managed care patients (85%) (P<.001 for pairwise comparisons). In multivariate analyses that adjusted for potentially confounding factors, FFS indemnity patients also had higher scores on the 10-item trust scale than salary (P<.001), capitated (P<.001), or FFS managed care patients (P<.01). The effects of payment method on patient trust were reduced when a measure based on patients' reports about physician behavior (eg, Does your physician take enough time to answer your questions?) was included in the regression analyses, but the differences remained statistically significant, except for the comparison between FFS managed care and FFS indemnity patients (P=.08). Patients' perceptions of how their physicians were paid were not independently associated with trust, but the 37.7% who said they did not know how their physicians were paid had higher levels of trust than other patients (P<.01). A total of 30.2% of patients were incorrect about their physicians' method of payment.

CONCLUSIONS

Most patients trusted their physicians, but FFS indemnity patients have higher levels of trust than salary, capitated, or FFS managed care patients. Patients' reports of physician behavior accounted for part of the variation in patients' trust in physicians who are paid differently. The impact of payment methods on patient trust may be mediated partly by physician behavior.

Physicians' experiences and beliefs regarding informal consultation

CONTEXT

Efforts to control medical expenses by emphasizing primary care and limiting specialty care may influence how physicians use informal or "curbside" consultation.

OBJECTIVE

To understand physicians' use of and beliefs about informal consultation.

DESIGN

Survey mailed in July 1997.

PARTICIPANTS

Of a random sample of Massachusetts general internists, pediatricians, cardiologists, orthopedic surgeons (n=300 each), and infectious disease specialists (n=200) surveyed, 1225 were eligible and 705 (58%) responded.

MAIN OUTCOME MEASURES

Self-reported use of and beliefs about informal consultation.

RESULTS

Generalist physicians requested more informal consultations than specialists (median, 3 vs 1 per week; P<.001) and were asked to provide fewer (2 vs 5 per week; P<.001). In multivariate analyses, physicians in a health maintenance organization, multispecialty group, or single-specialty group requested more informal consultations than those in solo practice (82%, 40%, and 28% more, respectively; all P<.001) and were more often asked to provide them (43%, 63%, and 14% more, respectively; all P<.05). Physicians with at least 30% of their income from capitation requested 38% more and were asked to provide 46% more informal consultations than those with little or no income from capitation (both P<.001). Generalists' overall approval of informal consultation was greater than specialists' (mean 5.9 vs 5.1 on a 7-point Likert scale; P<.001), and approval was strongly associated with beliefs about how informal consultation affects quality of care (P<.001).

CONCLUSIONS

Use of informal consultation is common, varies by specialty, practice setting, and capitation, and therefore may increase with current trends toward group practice and managed care. Because overall approval of informal consultation is strongly associated with beliefs about how it affects quality of care, this issue should be carefully considered by physicians who participate in informal consultation.

Intermediate TISS: a new Therapeutic Intervention Scoring System for non-ICU patients

OBJECTIVE

To modify the Therapeutic Intervention Scoring System (TISS) for intermediate and floor care nursing units.

DESIGN

Prospective study.

SETTING

University teaching hospital.

PATIENTS

In-patients on medical and surgical nursing units.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

Intermediate TISS retains 49 Original TISS items, adds 26 Intermediate TISS items, deletes 18 Original TISS items, and reweights 10 Original TISS items. Intermediate TISS score sums up the monitoring and therapeutic modalities used within the previous 24 hrs, each of which is assigned a weighted score from 1 to 4 points. A total of 3,073 patient days for 1,013 patients were simultaneously scored for Original TISS and Intermediate TISS over a 4-month period. A random sample of 435 patients was examined to identify those patients for whom Intermediate TISS added information over and above that obtained from Original TISS. Original TISS and Intermediate TISS correlated well, regression equation: Intermediate TISS = 4.4 + 1.4 x Original TISS (r2 = .83). We identified two groups of outliers, together constituting 96 of 435 patients: a) patients whose sum of new TISS items was unusually high relative to the sum of old TISS items included those patients with diabetes mellitus, those patients receiving serial electrocardiograms to rule out myocardial infarction and patients requiring cardiac or pulmonary therapy; b) those patients whose sum of new TISS items was low relative to the sum of old TISS items were largely surgical patients outside the intensive care unit (ICU), in whom Original TISS still worked well because it had been designed for surgical ICU patients.

CONCLUSIONS

Intermediate TISS is useful to score medical patients receiving intermediate or floor care in non-ICU nursing units within the hospital.

Combining census, dual-system, and evaluation study data to estimate population shares

"The 1990 [U.S.] census and Post-Enumeration Survey produced census and dual system estimates (DSE) of population by domain, together with an estimated sampling covariance matrix of the DSE. Estimates of the bias of the DSE were derived from various PES evaluation programs. Of the three sources, the unadjusted census is the least variable but is believed to be the most biased, the DSE is less biased but more variable, and the bias estimates may be regarded as unbiased but are the most variable. This article addresses methods for combining the census, the DSE, and bias estimates obtained from the evaluation programs to produce accurate estimates of population shares, as measured by weighted squared- or absolute-error loss functions applied to estimated population shares of domains."

Hierarchical logistic regression models for imputation of unresolved enumeration status in undercount estimation

"In this article we describe a logistic regression modeling approach for nonresponse in the [U.S.] Post-Enumeration Survey (PES) that has desirable theoretical properties and that has performed well in practice.... In the 1990 PES, interviews were not obtained from approximately 1.2% of households in the sample, and approximately 2.1% of the individuals in interviewed households were considered unresolved after follow-up....The missing binary enumeration statuses for these unresolved cases were replaced with probabilities estimated under a statistical model that incorporated covariate information observed for these cases. This article describes an approach to modeling missing binary outcomes when there are a large number of covariates."

Risk factors for pancreatic cellular injury after cardiopulmonary bypass

BACKGROUND

Pancreatitis is a known complication of cardiac surgery with cardiopulmonary bypass. Although ischemia is believed to be a factor, the cause of pancreatitis after cardiopulmonary bypass remains unknown.

METHODS

We prospectively studied 300 consecutive patients undergoing cardiac surgery with cardiopulmonary bypass. Serum amylase, pancreatic isoamylase, and serum lipase were measured on postoperative days 1,2,3,7, and 10. Pancreatic cellular injury was defined as the presence of hyperamylasemia (greater than 123 U per liter) with an increase in either the serum level of lipase (greater than 24 U per liter) or the peak level of pancreatic isoamylase. Trypsinogen-activation peptides, which indicate intrapancreatic enzyme activation, were measured in the urine of the last 101 patients studied.

RESULTS

Evidence of pancreatic cellular injury was detected in 80 patients (27 percent), of whom 23 had associated abdominal signs or symptoms and 3 had severe pancreatitis (2 with pancreatic abscess and 1 with necrotizing hemorrhagic pancreatitis). Two of 19 postoperative deaths were secondary to pancreatitis. In multivariate analyses, the development of pancreatic cellular injury was significantly associated with preoperative renal insufficiency, valve surgery, postoperative hypotension, and perioperative administration of calcium chloride. The administration of more than 800 mg of calcium chloride per square meter of body-surface area was an independent predictor of pancreatic cellular injury, and the increase in risk was dose-related. No differences were found in the level of trypsinogen-activation peptides between patients who had pancreatic cellular injury and those who did not.

CONCLUSIONS

Pancreatic cellular injury, as indicated by hyperamylasemia of pancreatic origin, is common after cardiac surgery. The administration of large doses of calcium chloride is an independent predictor of pancreatic cellular injury and may be a cause of it.